Hey guys, welcome everyone for the key different show today. It's a pretty exciting topic and I'm super thrilled for us getting started. We have some amazing speakers. Some of the speakers are good old friends of mine as well. So super thrilled about today. So one of the things we like to do on the show is to always
discuss about the whole topics on this impact of AI on blockchain here, but we can hash the same things as we always do or we can focus on those unique ideas that usually slip out of conversations and that's exactly what I like to think about and brainstorm and have this discussion right now and we're going to keep it casual. So if you do want to jump in, feel free to. You don't have to raise your hands and wait to be called. Feel free to unmute and
jump in. It's a casual conversation here and let's get started with the quick 10 seconds intro. I'll just give you a quick one for my end. I've been in the world of blockchain for 10 years in the world of marketing for 16 years and I'm the founder of Forward Protocol, a drag and drop no code solution for Web 3. I'm pretty excited to hear more from our speakers.
Hi, yes, good to be with everyone. Good to see a lot of familiar faces. I'm Andrew Lubon. I was a seven year professional athlete, did some angel investing in the space, was the COO of an NFT marketplace and now I was the first business development executive hire at Pulsar where the Google for NFTs. It's an AI powered search solution where you can find NFTs and we exist upstream of all of these
the marketplaces to make it a more user friendly experience and kind of try to do our part for the industry to make the UI UX simple and easy to use and extract a lot of the black blockchain application away from the user experience.
Hi everyone, my name is Anish Muhammed. I've been in cryptography for 20 plus years in crypto as we know it since for a while, like early advice at Ripple, one of the early members of the 3M Swamp team have worked on probably half a dozen layer ones, 20 plus protocols, working in ZK for the last seven, eight years.
On my third ZK startup, ML, Cloud ML, Big Data, old hat, wrote some of the early standards for cloud security alliance and other things, built an analytics platform for one of the largest banks in the UK. In the AI-ML intersection, as ethics of AI, I have a charity that I co-founded 10 years ago, used to teach occasionally at
GMU, mostly lecturing privacy, pursuing ML, and I've thought that that course is including ML for finance. Yeah, internally speaking, I've been advised, worked on a bunch of protocols like Ocean and stuff like that. So that's me.
Amazing, amazing. The next speaker, Edvin, why don't we start with you?
Hey, hello, Karina. How are you? Happy to be here. Yes, my name is Edvin Matan, CEO and co-founder of Brickin. We are a real world asset tokenization platform, been in the space since 2020, tokenized above 200 million worth, and now looking to explore new opportunities.
And thank you very much for here. Super excited to talk about how obviously machine learning AI is affecting us as a business, while we're looking in the space, and obviously provide some insights about what we're looking, how it seems for us.
Obviously, because the revolution, right, blockchain plus AI, it doesn't get any more sexy than that, right? Thank you very much for inviting me.
Yeah, yeah, I mean, it's perfect. This session has some pretty interesting speakers from different backgrounds, like cryptography, RWAs. This is going to be super exciting. Kathy, why don't we go with you next?
Sure, great. Thanks for having me. Key, actually, I think we've coincided in quite a few different events here and there.
Great to be on Twitter or X with you. Thanks for having me. And Edwin, I know Edwin. Well, I'm, I live in Spain as well. And we've coincided in events in conversation as well. So in great projects. So thanks for having me.
So I run Babs. Babs is a web three full stack agency. So we do everything from go to market all the way through to fully execution of any of that. So user acquisition, branding, UX, UI builds.
PR, I'm trying to think. We also, my co-founder, Nastya, who I don't believe is on today, she's at a birthday party, well deserved break.
Her background's product, so she leads all of our product team. So we can kind of perfect product or build product.
And I'm really excited to be on this today and have this conversation because I think 2024 is going to be the year where we see the projects that have survived.
The 2023 or 2024 rather be the year that have survived 2023 and we're already seeing a lot of interesting AI cases come to come to kind of fruition. So I'm interested to see how these this conversation is going to evolve. So thanks for having me and glad to be here.
Perfect, perfect, amazing. Why don't we go on with you, Sionis?
Hey, not sure if you can hear me. But I am, my name is Sal. I'm currently serving as the Senior Product Manager for NAI. NAI is basically just a social authenticity network that provides audible proof of social authenticity.
And we just leverage a Web2, Web3 and AI hybrid model with user-friendly account abstraction that allows basically your normies and no-coiners to kind of show up, join in, take part and, you know, it makes it a little bit easier for them.
We also have a DID piece which provides a global trust metric, which is a weighted metric using ML algos, as well as community voting, which can be shared across any other external platform. So like on Twitter X, I'm sure we've all noticed massive influx of bots.
So NAI kind of aims to mitigate that problem a little bit. Kind of head into the NAI Visualizer, audit the data yourself to see why they're authentic, have they just been sibbled, whatever else have you.
And yeah, that's basically it from NAI. Amazing, amazing. That's perfect. We have another familiar friend, Andrew. We have been on spaces together, spoken together before as well. So why don't you go next?
If you have trouble, I'm Andrew Lubon, a seven-year professional athlete, did some angel investing, was CEO of an NFT marketplace, Sports NFT marketplace, and then was the first business development executive hire at Pulsar.
Pulsar will during the Google search for NFTs. So kind of doing our part for the industry to abstract a lot of the technical and complicated stuff of the blockchain experience and make it super user-friendly.
The idea being you just use our AI-generated search to query like you would on Google, and it populates the results based on your preferences for anything that uses NFTs.
Perfect. And we have a last speaker, GBD protocol.
Hi, this is Peter here. I'm the founder of GBD protocol. We're a layer 2 blockchain that focuses on building a censorship-resistant AI network.
I previously founded a company called Jackson. It was really a pioneer in voice commerce technology.
We created an AI technology that was founded by Alexa and Google Home. So we signed partnerships with them in Java 5.
But I think there was one kind of caveat in web 2 is that there was a lot of gatekeepers and there was a lot of fragmentation and payments and merchant services.
So we started to move into web 3 where we could start to build AI in an unbridled way.
So yeah, right now we're focused on building a network for mining high-performance computing power for AI technologies and really being able to democratize that as opposed to how Bitcoin is.
So being able to provide open source AI developers a platform to develop on, having a centralization mechanism there, and then allow our compute power to really provide great computing for open source developers.
So I think right now a lot of it is going towards Silicon Valley, if you have billions of dollars, you can have the best AI.
But I think right now we're fostering an ecosystem and infrastructure for collective intelligence and federated learning.
So that's what we're doing right now, and I was surprised to see the $50,000 worth of funding continuing to take in more capital right now and launch on centralization.
Yeah, very exciting time for AI and blockchain.
Amazing, amazing. Well, thank you so much for that. And now that we have introductions out of the way, I just have one quick breakfast for everyone here.
There's a pinned tweet right about here. If you can tag your friends, retweet, like it, we can get the heat turned up a little more as well.
And as we get on, I want to set the context here.
AI, as it came into the space, we could see a few things happening here.
AI did impact positively and negatively to the world of blockchain. On the negative side, we know that a lot of venture capital money has moved on the AI side.
And there's been a lot of buzz and attention going there, all of that.
But on the positive side, I see that all the AI businesses out there, irrespective of who they are, what they are, they all do need the source data, and data is not going to stay free forever.
So when that happens, the data needs to be monetized, and I believe blockchain is going to be that beautiful solution out there.
But beyond that, what I'm keen to understand is what other impact do we see? What are the other areas where positively or negatively AI is impacting blockchain?
And why don't we start with Anish? I know you have a very deep technical understanding and expertise on the blockchain space, and I would be very, very keen to hear how you see this.
Thank you very much. Let me add a bit of caveat to this thing. So, you know, nobody gets surprised. So I am one of the original members of the Ethereum Swamp Team.
So last of my Swamp Summit Talkers in 2018, I pointed out some of the challenges we have to actually do what you are suggesting.
So what you need to have is if you want to do price discovery. Price discovery requires that the cost of doing a transaction, cost of storage, and the speed should all align.
So effectively what that means is like if you want to actually have some data stored somewhere, you want to do some mechanism of price discovery, and then go and use it, that becomes a bit more challenging.
Things have improved. 2018, in the three terms, is a very long while ago. So for those who don't know what Swamp is, Ethereum had three pieces.
One was EVM, which everybody knows. Then there was another thing called Visper, which was inter-process communication.
And the third one was Swarm, essentially distributed storage. So that's what I was talking about.
It was Ethereum's attempt at making sure we have distributed storage, and then use the storage to really monetize for various things. So that's the start of the thing.
So going back to your question of will it be possible for us to actually store data, be used, blockchain as a mechanism for fairness and price discovery.
The first part, I think, is a challenge. So you can look at all the different projects, protocols, as we described them. So whether it's Filecoin or whether it's Swarm or other, you know, a lot of the RBs.
All of them, as far as I can see, have run into this fundamental problem. The fundamental problem we have is like compute has been growing on a different curve as to network.
It's a network bandwidth in that sense. So if you have been looking at compute and a creation of data, the creation of data has really gone almost exponential or non-linear for sure.
But if you look at the overall bandwidth we actually have, that's more or less linear. So what we end up is actually having this real challenge of having lots and lots of data and having to do transfers pretty implicitly.
So one of the key challenges that are people who play with LLMs is like the real data size is that you need to play around to do various things in any way, shape or form to be distributed.
So there's an arbitrage to be had to do any kind of decentralized learning mechanism of any kind. And then there's a real question of what I call geographical arbitrage for storage and network.
What I mean by this is like if you have, say, 10 petabytes of data in, say, Africa and 10 petabytes of data in East Coast or West Coast of the US, the real cost in that sense is very different.
But normally all protocols don't see this difference. And most of the things that are there in the Web3 world don't see the difference.
So my one-line summary is, yes, we could definitely have blockchain as a mechanism to do price discovery, blockchain as a mechanism to a degree verifiable compute.
We could actually have provenance of data, provenance of data in machine learning pipelines or data engineering pipelines and state transition in machine learning models.
But will that be economical? Will it be truly feasible at the throughput we want? I am really not sure.
Well, that's definitely a good perspective. And I believe we do have time for that.
I don't think the AI industry and the monetization of data is as immediate as we foresee. I think it's still at least going to take a year or so.
And by the time, I think the progress in the blockchain space would also be caught up to that.
OK, do we have answers or feedback from different angle? Like probably, Katty, do you have something to say from marketing?
Because your background and what you see there is, I believe, is an opposite to what Anish, for example, from a technical side would be seeing.
So I'm keen to hear your thoughts as well on this.
Meanwhile, if anyone else has any insights, feel free to unmute yourself and jump in as well.
Me personally, I do believe that data is the key factor that's going to contribute.
And despite the challenges, I think when there is a problem presented, that's when the solution can come.
And that's when more value is built and more opportunity or more market interest is being capitalized on that particular sector,
which would force startups to focus on that area to innovate, to come up with newer solutions.
And I think the curve at which we in the blockchain space have been developing and growing, I think it's been incredible, especially in the last two years.
So I'm very, very, very positive on that.
If anybody has thoughts, jump in. Otherwise, I have our next interesting question, which deals with the projection.
So, AI at this point in time is obviously buzzing and we also are seeing the bubble grow, but there are challenges that we already see with the trend.
There's a time when the interest could shift away from AI and focus back again because we are in a positive market right now.
I don't know if it's a bull market or not, but we are in a positive market in the blockchain space.
So now when this heats up as well, there's going to be a shift of interest.
It's going to be a point where the money has to flow more into this hot, timely opportunity.
So what do you guys project as the timeline towards that or when you see that happening?
I'll jump in here and I think it becomes a little bit ephemeral when we start thinking about the changes between what the financial interest is.
So I think in crypto, sometimes we call it the hot ball of money and people are usually following whatever's trending within the AI space.
I think it less becomes things fall out of vogue and a lot of the AI tech that people are kind of enamored by right now.
They stop paying attention to it and I think it more so becomes almost a regular part of your life.
We've already kind of experienced this from last year.
Once a chat GPT kind of became almost ubiquitous in terms of what people are using,
a lot of people started using it almost like a chat version of Google, which is fine.
It's a use case, but as that starts to become a little bit more normalized and as AI consumer products start becoming a little bit more sophisticated
and people start using them and integrating them more in their day-to-day lives,
I think it's not so much that the attention of the hot ball of money leaves AI,
but more so that AI starts supplementing that hot ball of money to kind of hit the things that are going to start becoming ephemerally trendy.
Does that kind of make sense?
I was just going to be like a bit, how could I say, contrarian.
Just full disclosure, I've known Imad, I've known Mario.
I've been in the space long enough to have known the people that you probably recognize.
Some of them might describe me as their friend and I've understood what they do and how they do.
And I've also been on the other side helping the so-called F2 VCs understand what this is.
And the way I think about it as the money that's required in the AI space to actually move the envelope in a significant sense is hundreds of millions and billions.
And in the normal investment cycle in F3, again, forgive me for saying this, we haven't reached that level yet.
And the real, real, real bottleneck is access to hardware.
So if you can get hold of Jensen and get him to give you his supply of whatever that is, your startup takes off for 100%.
So that level of binding to access to fab and access to hardware acceleration in the form of NVIDIA's products or any products other than rare instances of Bitcoin mining kind of thing,
To me, so the two threats are separate in one sense of speaking.
You know, the investment that went into anthropic or stability AI would actually start up maybe tens of hundreds of Web 3 startups.
That's kind of how I think. Maybe I'm wrong.
I see or I'm going to definitely buy that. Yes, go ahead. Go ahead.
Yeah, I definitely agree with you. I definitely agree with you, Anish.
I think that, you know, there's a lot of money that's flowing around and it's all concentrated at the top.
And I think with Web 3, we're going to be able to democratize that and allow for open building and open development and allow for participants in the network to start to stake in their favorite AI model.
You know, them and building Web 3.
I do think that we kind of have to decouple a little bit of, you know, the traditional Web 3 where it's, you know, it started to just lay over or just create new financial products.
Right. But I think when we look at, you know, the future of blockchain and AI, it's really that we're solving a problem here and we're using blockchain to do it.
And then there's obviously an incentive mechanism to it or what's important, but it's not a financial product first.
I think what's important is that it becomes an asset and a commodity where AI compute power starts to become a commodity.
AI data starts to become a commodity. It starts to become a better store of value for AI tokens. Right.
I think the common criticism of Bitcoin is that it really doesn't have a store of value. Right.
But when we start to think about, you know, you know, Google has something called proof of resources. Right.
So as they start to provide high performance computing power and AI compute power to these open source developers for AI, it really allows it really allows us to take all of that data that we collect and utilize the energy that's being
spent to mine, high performance computing power versus just like doing proof of worth of money, you know, Bitcoin, for example. Right.
So I think we're we're reaching a new maturation cycle in roughly where I think that AI is an exponential technology and I think it will create prolonged and lasting rise to AI and have more stability or a prolonged and lasting
rise in value to the web3 community and web3 in general because it creates value in terms of taking that AI data and making it a commodity.
So like, for example, with our product, it's really we were utilizing polygon CBK and utilizing the Libyans for option AI data processing.
And that can also, you know, decay proofs and then take AI data and prove and create a trust layer for AI.
So AI is to me with AI or AI is to me with human. There is that verification layer that provides a trust to AI that we don't see in web2 today.
And I think that inherently build a value. And then when you start to do that and you start to store like prompts and responsive on chain, that allows us to build a collective intelligence.
And that can give rise to like a super intelligence, right? If everyone's building on the open, providing open data sets that are verified and we're seeing new things emerge in terms of content creation from AI.
I think there is an opportunity to really evolve the space into much more than just finance, right?
I think it's really, you know, how finance embedded into existing infrastructure of what we're trying to build, which is I believe that the next iteration of the internet will be completely generated by AI.
And that's what we're looking to do with the Liby protocol to really build that next AI internet, like censorship is in some areas of the internet today.
Hey, so also jumping off that, I think something that hasn't really been talked about, I mean, since I joined was more so the ethical angle of AI and where the industry started, you know, when it really hit its hype cycle and not fall of 2022.
And since we've really broken the bounds of obviously the tech consumers Gen Z or consumers are starting to use it within their school systems and so forth, college institutions are starting to use it and so forth.
There's still a lot of active legal regulation that's still happening in real time, and it can inevitably go off the rails with, you know, the wrong programming and so forth.
So I think similar to the crypto space and what we saw the last few years, obviously, it's going through its hype cycle.
But when we produce an event at New York Tech Week last October, I think it was October, November, I can't even remember anymore, like 80% of the events through a 16 Z's calendar were AI related.
So that said, do with that information as you will, I don't think it's slowing down anytime soon. I do think obviously certain regulators, lawmakers and so forth will jump in where necessary but
to be honest, when you look up any regular AI news resource that comes out on YouTube on a day to day basis, like there's already 20 to 30,000 views on some of these top creators that are coming down, coming out in within 24 hours and they're real views.
It's not some, you know, arbitrage. They just bought views. These are real builders that are watching the tools that are coming out on GitHub as well as YouTube.
So, I think there's a lot of different possibilities that come along with it and it's not slowing down from what I've seen. So,
Sorry, Kia, I was trying to talk before but my internet has been horrible because we're having storms. I apologize. You asked me a question I couldn't connect.
I just, I just wanted to add a little bit on to what Lina was saying in the sense of from like the marketing and branding perspective, that it's not slowing down and there's also like this fine line like you said in terms of ethical kind of like what is proprietary at this point.
Or if we're creating a campaign and we're doing content or we're talking about insights and knowledge and if you're using the majority of the kind of the source is AI, does that then become like your IP? Are those your thoughts?
Are you actually like is that your brand identity or is the AI producing your brand identity? So, I think that's one of the things we're playing with right now.
I was listening to I guess what's happening right now with chat GPT and this fact that they're having this whole lawsuit with the New York Times because it turns out that like something like the majority of the algorithm is being was pulled from the New York Times and they didn't have any of the licensing or any of the agreements, whatever that might be legally to be able to use those sources to pump the algorithm.
So then it becomes like so like for me as a marketing leader, it's you kind of are in this position of like whose thoughts are we creating? Are these our own? Are they someone else's and do we have permission to be using some of these thoughts without necessarily sort of like citing the source?
So, it becomes super complex and I think that people just as you were kind of saying, Lena, like people kind of lean on this whole AI thing and it's like, oh, we'll just use it for in any capacity without understanding that these regulations will change.
Things will change over time. We need to be able to be cognizant of it and be able to pivot and to be to stay honest and transparent and like the work that we're doing, I think.
It's a little off what we were talking about, but I think it's an important part of the AI discussion.
Yeah, I mean, I was just going to be a bit controversial. So, full disclosure, I am the co-founder of EthicsNet.
It is a charity set up in the UK in existence for 10 years. We got some funding from the government of Netherlands to help look at research on ethics of AI.
So, very familiar with a lot of the challenges. I think a part of the thing that we are really recognizing are set of challenges that have actually impacted our society already.
There is this construct called surveillance capitalism, Susan Kampana is her name.
She wrote a book in 2014 and that's when this recognition happened. And then what has happened is then that shit and election of Trump happened in the US.
Cambridge Analytica came in, that's use of data science on public data.
Now what we are kind of saying is like, okay, we have ML models, we have better ML models.
So, we have models that actually kind of learn by themselves.
And all of this has open questions. I kind of mentioned in brief that are kind of three things of authenticity of data, the authenticity of data engineering pipeline, three in one sense is describing.
I could describe this as explainability in one sense or the way in which the machine learning model transforms.
What is happening mostly right now is like we have a black box, you throw some stuff out of into it, and something comes out of the other end, we make a bunch of assumptions, and then we go, okay, these are the assumptions that we're going to work on, we will see if it works.
And so far, as the previous speaker was mentioning, or the previous speaker was mentioning, the legal framework that's there is really, really not mature.
So again, full disclosure, a few years ago, WVF, economic forum, they had two sets of workshops on impact of AI on finance, and I was one of the participants, the same set of questions that we are kind of discussing came up there.
My view was, and still remains the same, a lot of pieces are still unexplored, and understood, and the cost of the society and the trade offs are not understood either.
So, you know, cost of society for application of political data science was made very, very clear by Brexit, and to a large degree made very clear by all the successes that Cambridge Analytica brought to various parties around the world.
So just imagine this, change Cambridge Analytica with AI ML model, name it whatever you want, and see the impact that will have on our day to day lives.
That's the start of the game. You can add that, you know, you can add layers and layers and dimensions to it, and you can see how the world would actually be in a few years time.
Perfect. I don't want to bring the focus back to some of the core aspects of how AI impacts blockchain side, more specifically, how the AI algorithms revolutionize the smart contract functionality.
From some of the areas that I'm looking at, I see that even in Forward Protocol, for example, we do have AI doing its security audits on smart contracts that are being uploaded or helping people with describing or explaining what's in the smart contract.
So, but those are the basic parts, but I'm more keen to share your thoughts on this.
Yeah, I would like to jump there on that conversation. So, as we know, like chat GPT or range or whatever platform we use is supposed to help us, right? It's not supposed to replace us.
If we use it as a platform that can amplify that work, then we become more efficient.
So from the calling perspective, for example, here in brick and all of our tech team has access to Compilot on GitHub, because it really helps them to not replace what they have an idea, but also amplify their scope,
or how can they optimize the coding abilities. So that kind of like optimizes the work because they can perform better.
But it's not supposed to be fully replaceable yet. I mean, even though the systems are getting upgraded on a monthly basis or whatever, I mean, we still have to be secure that some of the things that our coders think or want to develop, it still may make them.
So by us providing these tools, it's just how they get to be better. And it's kind of like they have a pioneer that they can really code with, but it's never supposed to be replaced.
And also from our audited team internally, it has also been a tool for them to optimize how they work and see whether they have the penetration tests or what they should be looking into.
So, for example, here in brick and we just AI across departments, just because it's a tool. So, for example, I'm a lawyer.
My working tool, it's always kind of like work and what is very rudimentary. I just write.
But if I amplify it with Charlie PT, it's like, hey, can you tell me if my class is correct?
Where should I be looking or where is my default or what is my whatever? My contracts have actually become so much better or my teams contracts just because it's just an active tool that we allow there to be there.
And but I still strongly believe that we should not replace because if you throw it to everything, everything just becomes the same.
And I'm talking with the marketing. I'm here on what is concept. It is very easy to now see if you actively using these tools.
Who is just shooting requests to charge it? And it's like you sound the same as the hundred others post tweets, even websites now, like it's like it's the same.
Your USB and another people's USB. So, I mean, there's going to be a point in time where everything is just the same.
So the radical difference is going to be who understands the use of these tools to become better than one that just becomes a sheet to be exact same.
So I think that's going to be crucial in the near future.
So I'm going to be contrary in here again. So just I'm going to give you a bunch of data points and dry my logic. So if you know, imagine it.
So imagine this is the way by which you used to measure machine learning algorithms for image recognition.
So a few years back, it just got better than humans. Right. I think human error rate is like six percent.
Most of the models right now, what we have is, you know, one one point five percent kind of the thing.
So they definitely the tools have reached a point where it could do a job better than an average human.
And the other part I wanted to say is certain use cases are really, really, really useful.
So let's think about the blockchain use case. Blockchain has a unique property.
The thing I want to say is like your mempool is available. Your state spaces are there.
All your smart contracts are there. All known vulnerabilities are there. This is a very, very unique space.
So if you were to train something, your ability to get results are incredibly high.
Whereas everything else, all the state transitions and everything else is not that visible.
So in my mind, there is a possibility that some of the AI tools and mechanisms might easily outperform most of the average court auditors
or average, you know, what I call workers of any kind. And then this is the real challenge.
It's like, you know, how would we actually make sure that as a sense of fairness, sense of value creation, distribution of value creation,
all those things are kind of done in a manner where society could still survive.
That to me is like the bigger problem in my mind. Maybe I'm really cynical. That's what I think.
That's a good perspective there. But just to touch base again on the other aspect of it, the scalability aspect.
Do you think there is even anything closely related where AI can contribute to solving our scalability challenge or energy consumption issues that we have with the blockchain networks?
Do you see the trilemma of blockchain being positively helped by AI through any means?
I mean, at least from my end, the way I foresee is that even though it may or may not be a direct contribution,
each of these parts, the three areas of the blockchain trilemma, they all do need a lot of effort, a lot of products, a lot of creation in its own way to solve it.
So it is something that has to be improved upon, built upon, maybe through various products and startups or through further innovation.
And I do see AI playing a role across board. Like one of the speakers rightly said, AI is not sure to replace us, but to support us.
And it's very true. And that support is going to be crucial for us across this across board to solve these challenges.
But if there is anything specific like an innovation or something that you see on the horizon, I'd be very, very keen to hear.
Because the only thing I hear, or at least I heard in the past, is all the scary stories about how quantum is going to derail blockchain when it comes or when it's ready.
But I think AI, on the other hand, can help in more than one ways. And yeah, open for your thoughts.
I think maybe one positive use case is crypto or blockchain needs a token. The token is digital money.
AI is going to revolutionize content. One easy piece is using blockchain to legitimize content that's user-generated versus AI-generated or some combination therein.
And it's a way of proving who's creating what and validating who's creating what.
I think that's a strong use case that the tokens from the blockchain when married with AI can pass on to the end user.
Well, a funny context to what you said, the AI and human-created content.
In the past, when you have a need, you create a job post and you post out there, and you have applicants applying.
And from the way they write, you know what is their intention, what's in their mind, what do they prioritize.
All of that, it's the stale that they write, this quality of their communication. Everything can be figured out from there.
But right now, you put up a job post, which you generated with chat GPT, and then they reply back, and all the replies are templated or all the replies are generated.
You just don't understand. It's almost like AI talking to AI through all of us. It's a manifestation through us.
And it's just funny. It kind of makes certain things funny and challenging.
But yes, I believe there's always a value with identifying if there is a way to identify human-created content.
Although I see data coming to blockchain, it's happening after the content is created.
So whether it's human or AI generated, it reaches blockchain as a second step.
And I'm not sure if there is any way for blockchain to identify and isolate it.
You do already have certain algorithms that can do that at this point.
I'm not sure of using the words and patterns, but I don't see blockchain directly impacting or contributing to that part.
If I may, I wanted to just say something about something you referred to previously, just to be sure that everybody's on the same page.
So when everybody talks about quantum computing, and we talk about the challenges blockchain is going to have,
the assumption that everybody's making is a following.
That in the next few years, our public key cryptosystems that we are currently using,
which is currently based on the ECT or the elliptical discrete law problem, essentially using elliptic curves, will remain to be the same.
There are a bunch of public key cryptographic primitives that NIST has actually recommended that are quantum resistant.
So in the quantum computing world, there's an algorithm called Shor's algorithm.
So all these so-called hard problems become linear problems in that sense.
Then going back to the difficulty in recognizing human-generated versus machine-generated,
if you just think in a very simple sense, think of a GAN, you know, the Analyze Opportunity Network.
What it's trying to do is simply put like having a discriminator and recognizing between the real thing and the thing that's been generated.
To a degree, that is incredibly, it's going to be incredibly hard to actually recognize what is real, right?
And this is going to only get worse. As the models improve, you can think of the two paradigms that are there on both sides of the fence.
One side of the fence is something.
So the basic construct that actually allows us to keep track of things is hash functions.
And one of the properties of hash functions is this, you know, collision resistance.
And the other one is like the change.
It actually that really happens when you just flip one bit.
OK, so it's a propagative thing.
And the inverse of that is kind of what machine learning does.
So in my kind of a philosophical hat, I see two competing paradigms competing against each other.
And I don't know which one will win and which one will.
You'll see the effect. I mean, we'll see the effect of it.
That makes sense. So with that being the case, when I'm keen to understand if I can be a game changer with the frauds in the blockchain space, because in the crypto space, there are a lot of different scams, rock pools, a lot of different things going on.
And we do have patterns behind it, patterns of wallets or patterns of particular kind of behavior, etc.
And I believe since it's data and it's an on-chain data, AI can contribute a lot to detecting or deducing to what is the risk of potential fraud something might be.
I kind of already like what I see with Genosis safe. Oh, yeah, go ahead. Go ahead.
So I was about to say something. Again, the thing I would say is like this is nothing new.
You had credit cards when credit cards came in, and even you can go to PayPal.
When PayPal started, there was a whole bunch of frauds. When the fraud in the PayPal ecosystem was high, they created a company called Palantir to actually apply ML, essentially ML and statistics.
So ML is like a subset of statistics, truly speaking.
So to find out fraud, and then you can see how much it has grown.
So the same thing can be applied to projects in blockchain.
And I'm pretty sure you could have a pretty reasonable outcome.
But at the same time, I'm pretty sure you've had your credit card and you went and tried to buy something, and it got rejected.
And the pain it brings to you. That's a false possibility.
So this is a question to be had.
What if you, being a silent contributor, you don't do marketing, so your web presence is very minimal.
Your name is on a Genesis block. You are the person whose name is on a couple of white papers, all on papers, whatever the crap is.
But the fact that you are never being a marketer works against you because ML doesn't know all the other things that are happening around the world,
but recognizes, because I'm sure the people that are doing marketing recognizes the fight between Google and SEO guys.
So this is literally the same thing.
All the people who actually are interested in creating projects that want to scam people, the moment you have an ML mechanism,
you will be able to improve the detection.
So if you go back and look at, say, Visa or MasterCard, I suspect they have like, you know, two, two, three, four, five, six percent of whatever fraud happening on a consistent basis.
So in spite of LLMs, in spite of GANs, in spite of everything, the reason being, you know, is AI versus AI.
The fraudsters will start using machine learning models to fight the people who are trying to fight it, right?
And defending is harder than attack.
And I think anything to do with analytics, anything to help prevent or warn is coming soon.
I can almost see them happening this year in 2024 with various type of apps.
But of course, their accuracy.
It comes with denied transactions or blocked PayPal account with the balances.
I guess many of us have always faced it.
But I think there is already a few DEXs and Safe and others having some form of risk assessment as part of their transactional information, which is good.
And if that can be powered by AI, especially when people are jumping in or aping into various things out there,
I think it can definitely save or at least get people to pass for a second.
Of course, it might not be accurate.
Of course, I think that's a part I like with Genesis.
Even if you're sending money to a blank wallet, a new wallet, it actually warns that it's a new wallet and it could be wrong.
Most likely, it is not wrong.
It's just a new wallet, but at least that's a clear warning.
And that will solve the challenges that you mentioned, Anishan.
A person who's not a marketer trying to put a product out and that gets warned.
But I believe that warning, if it's phrase smart phrase right across board, I think it can save a lot of money for people in the crypto space.
And I see that over and over, even with things like Twitter tweets and stuff, wrong accounts trying to tweet something impersonating and people still lose money every day.
And I believe that there's a big product market fit on more than one industry around AI, helping people warn and keep them on the safer side.
Yeah, so I was going to add something, but yeah, I definitely agree with you there.
And especially with the advent of AI technology and just getting more and more intelligent with generative AI.
I think the traditional ways that we're doing things in web 2 or the traditional financial infrastructure is actually more susceptible and vulnerable to AI.
Like it could just change, generative AI can change a whole bunch of things in the database, right?
So I think as we start to move towards more organizations becoming, moving from traditional organizations and autonomous organizations,
I think the blockchain becomes rather crucial to the trust that we've all lost from some of these centralized companies.
Because that data is extremely vulnerable and especially with like, because we work closely, our company worked closely with some of these technologies that are working on off-chain data processing for AI using ZK proofs and like ZK rollups and things like that.
So it really, the implementation of ZK, especially ZK proofs, and I would say with a polygonal palladium is that it's able to prove data off-chain, especially with AI processes.
So creating that trust layer is going to be really important for preventing any type of fraud.
But even like you said, even when it comes to phishing, right?
I think that's not just like happy and breaking and crutches.
I think if even just phishing as well, we can have a trail of what was set to the AI and what did the AI, you know, what was set to the AI, what did the AI respond with?
And that's proved and it's immutable using ZK proofs.
So I think that in itself is what's going to add a tremendous amount of value to the upcoming, what, three ecosystem.
And I think we're crossing into like what for territory at this point, you know.
But I think it's really important that we start to look at a lot of these use cases and technologies that are being developed and really jump on that because they're really about the create that can bring an explosion of the AI technology that will be on-chain.
And I think what's really good about AI technology too, even from like a doomer standpoint is, you know, if we put AI on the blockchain, if it starts to become an existential threat to human beings, we can just defund it, right?
Like I think that's like the beauty of the blockchain.
So it puts people more in control of their future.
And it actually allows us to innovate more without fear, once it starts to move on to the type of infrastructure, these types of rails.
Yeah, I think without appreciating the challenges that the Web2 world has in terms of integrating a lot of this stuff, I think the promise is that AI plus blockchain provided that companies in the space, evangelists, et cetera,
really educate the populace on how to use and read blockchain.
I think it solves a lot of the concepts of what is legitimate, what's not.
I think that's a big challenge.
I think we're still very far away from people of the average user understanding blockchain in a meaningful way.
So in some instances, you have to be worried that we're abstracted away from all of us who may be in bubbles, whether AI or crypto, and then you marry the two and you're very far away from your kind of average user.
But I think they still are on track to solve those problems.
And I think it's on a lot of us in the space to do the education or create UI, UX, that abstracts, that experience way to a way that communicates how to read these things in a meaningful way for the casual user.
There's this one risk that I've been thinking about.
AI is hungry to learn, and it's always learning something and analyzing, storing, comprehending information, and blockchain is immutable.
So just we put some information out there.
Do we even see any risk here or what are your thoughts?
Me personally, I don't see a challenge with it because, of course, the nature of blockchain is permanency, but that's great.
And unless until the data is confidential by itself, I don't see that to be a problem.
And it's something that I think can only help more from a big data standpoint or data analysis and interpretation.
But I don't really see that hurting or creating new risks in any form except for the fact that if it's a public blockchain and the data is publicly out there, it's not like you could stop someone from seeing it or reading it.
If we decide to change it, like Twitter and One Find It decided, OK, they don't like the data being read and they started blocking the bots.
But when it comes to blockchain, do we block it at explorer level?
Someone can run a node and read every information out there still.
So it's something that people should rather think about the future implications of it before putting the data on chain outside.
Yeah, I definitely agree with you.
Like all of the data on chain, you know, I think it really all depends because there's also like privacy of the user.
You know, as your platform.
So I think that's why I like off chain data processing and using things like the big groups allow for that privacy as well.
So it doesn't necessarily put all of your personal information on the chain.
But I think there's some things that require or that would be really good for the A.I.
Especially anonymized data.
If data is anonymized and provided openly, it allows for faster and better innovation and trusted innovation at scale.
So I think that there are there are tools out there or there are ways to architect the product to be both private and public at the same time.
And I think with A.I. data, it should be that way.
But I think, like, I think being able to put the A.I. data on chain allows it to be easily shared.
And then also putting it on chain helps it just be more trusted by the user in general.
So and then it also allows for a federated learning to occur pretty quickly as well.
So, yeah, I definitely think that this type of data that we collect from our products shouldn't just be siloed into like chat or open A.I.
Like, I think it should be open.
And I think, like, you know, I think a lot of times, like, you look to like, we just give out our data.
And like these two companies just take all of our data and they're just out of you.
But like we described a new way of thinking where everyone has the right of ownership to their A.I. to A.I. data.
And that starts to become a commodity in itself.
So it's just another way of looking at it where it's just like, OK, not just the, you know, that one company worth trillions.
Like now we all have an incentive to utilize this A.I. work, provide A.I. data to the A.I. together.
You start to get rewarded.
I think that's going to be a pretty interesting new space in the future.
Just a quick shout out here.
We do have a space the same time every, you know, every week this day on Tuesday.
So next Tuesday, feel free to join us.
Feel free to participate here as well.
We have a pretty interesting topic scheduled for them.
But in the meantime, thank you for every speaker, everyone here who contributed their valuable thoughts and all the listeners who are part of the community right now.
Thank you for all the like shares and your inputs.
Super excited and I strongly do believe that although A.I. poses its own sort of challenges to the space, it also brings a lot of great value and they both have bigger markets to individually grow and sustain.
And I'm excited to see all of that happen in our lifetime.
I think we are all pretty blessed to be in this time right now when all these major trends are converging and coming into life.
So thank you, everyone, and look forward to seeing you all next time next week at the same time. Thank you.